The Value of Three-Dimensional Printing Spine Model in Severe Spine Deformity Correction Surgery

Study Design: Retrospective case-control study. Objective: We aimed to evaluate the value of 3-dimensional printing (3DP) spine model in the surgical treatment of severe spinal deformity since the prosperous development of 3DP technology. Methods: Severe scoliosis or hyper-kyphosis patients underwent posterior fixation and fusion surgery using the 3DP spine models were reviewed (3DP group). Spinal deformity surgeries operated by free-hand screw implantation during the same period were selected as the control group after propensity score matching (PSM). The correction rate, pedicle screw accuracy, and complications were analyzed. Class A and B screws were defined as accurate according to Gertzbein and Robbins criteria. Results: 35 patients were enrolled in the 3DP group and 35 matched cases were included in the control group. The perioperative baseline data and deformity correction rate were similar between both groups (P > .05). However, the operation time and blood loss were significantly less in the 3DP group (296.14 ± 66.18 min vs. 329.43 ± 67.16 min, 711.43 ± 552.28 mL vs. 1322.29 ± 828.23 mL, P < .05). More three-column osteotomies (Grade 3-6) were performed in the 3DP group (30/35, 85.7% vs. 21/35, 60.0%. P = .016). The screw placement accuracy was significantly higher in the 3DP group (422/582, 72.51% vs. 397/575, 69.04%. P = .024). The screw misplacement related complication rate was significantly higher in the free-hand group (6/35 vs. 1/35, P = .046). Conclusions: The study provided solid evidence that 3DP spine models can enhance surgeons’ confidence in performing higher grade osteotomies and improve the safety and efficiency in severe spine deformity correction surgery. 3D printing technology has a good prospect in spinal deformity surgery.


Background
Medical 3-dimensional printing (3DP) is an emerging technology and a powerful tool for recreating details of human anatomy. The 3DP and computer-based simulation are becoming a routine part of pre-operative planning, intra-operative guidance and patient-specific implant modeling in many centers across the world. 1,2 Severe spinal deformity correction is quite challenging for spine surgeons because of the complex and variable anatomical spine structure. Although spine computed tomography (CT) scan reconstruction can provide visual 3D reconstruction for doctors on the screen, the guidance of traditional medical imaging for the preoperative surgical planning and intraoperative operation is still limited. The 3DP technology based on the CT scan data can present surgeons a solid spine model in its original proportions with high precision. It not only plays an important role in teaching medical students, consulting with patients, and making surgical plans, but also assisting surgeons intraoperatively and shortening the operative time by visualizing the anatomy of organs and tissues in some recent investigations. [3][4][5] In this study, we aimed to analyze whether the printed model could indeed assist in the surgical treatment of severe and complex spinal deformity by reviewing the cases at our center since the beginning of the 3DP spine model application. We hope that our study will help to optimize clinical decision-making and clarify the clinical indications of 3DP technology.

Methods
Ahead of the study, we received the approval of the ethics committee of BJCY hospital, CCMU (Approval number: 2017-2-21-5), and it was performed in accordance with the Helsinki Declaration of 1964, and its later amendments. Informed consent was obtained from all individual participants included in the study. Patients diagnosed with scoliosis, hyperkyphosis or kypho-scoliosis were retrospectively reviewed following posterior pedicle screw fixation and fusion surgery. The inclusion criteria were adult scoliosis, congenital scoliosis, neuro-fibromatosis scoliosis, neuromuscular scoliosis, and scoliotic Cobb angle !80 or kyphotic angle !80 . The 3DP spine model was made voluntarily. Surgeries performed with the guidance of 3DP spine model were defined as the 3DP group. Surgeries performed by free-hand, without the guidance of 3D model or any other navigation techniques, were defined as freehand group. All of the cases were selected from the same center's database during the same period performed by 1 experienced surgeon (Y H).
Propensity score matching (PSM) was used to compare and analyze between groups in order to eliminate confounding factors and keep the baseline data of both groups consistent. Cases in the freehand group were retrospective inconsecutive matching cohort. Group characteristics were matched after one-to-one PSM according to parameters including age, gender, diagnosis, and Cobb angle of scoliosis or kyphosis.
The spine model was printed at the original 1:1 ratio by 3DP technology (BF-SLA-600, Shanghai Blackflame Medical Technology Co., Ltd.) based on the CT scan data (DICOM format data from Siemens CT machine, SOMATOM Sensation 16, Siemens AG, Forchheim Germany) of the whole spine. Resin was used as the printing material. The print error of the model was < 0.2% per 10 cm.
The spine model can be used to assist in preoperative evaluation and surgical correction, to improve the surgical plan quality, especially for increasing the accuracy of screw placement at the congenital dysplastic vertebrae and the accuracy of vertebrae resectomy. For intraoperative screws placement and correction operation, the 3DP model was carried out as the role of the materialized 3D CT scan reconstructed picture to aid the intraoperative workflow in a number of ways. After the screws were placed, the C-arm fluoroscopy was used to check the position of the screws by 2-dementional images, no advanced navigation tools were used during operation in both groups.
Patient clinical information, surgical data and radiographic parameters, including age, gender, diagnosis, operation time, blood loss, fusion levels, screw density, complications, Cobb angle and correction rate were collected and analyzed. All radiological parameters were measured by 2 experienced doctors independently and the average value was taken as the final result. With regard to screw position classification, a third senior reviewer will evaluate the screws with controversial positions to determine their final classification.
All patients underwent postoperative CT examination of the spine to check the accuracy of every pedicle screw. Gertzbein and Robbins standard was used to classify the screws and define the screw accuracy ( Figure 1). Grade A: the screw is totally inside the pedicle and does not penetrate the cortical bone of pedicle. Grade B: the screw is inserted within 2 mm or less penetration beyond the pedicle. Grade C: the screw has a penetration of 2 to 4 mm. Grade D: the screw has a penetration of 4 to 6 mm. Grade E: the penetration is > 6 mm. Grade A and B screws were defined as the accurate screws. Internal fixation related complications, such as nerve root injury, spinal cord injury, cerebrospinal fluid (CSF) leakage were recorded and analyzed.
The SPSS version 18 software (IBM Corp., Armonk, NY) was used to perform statistical analyzes. Two-sample independent T test was conducted to assess the differences of continuous variables with parametric data between the 2 cohorts. Fisher's exact test was used to analyze differences of categorical variables in outcome variables, where P value of .05 was considered statistical significant.

Results
Thirty-five spine models were printed to assist the surgery in 35 spine deformity patients, including 10 men and 25 women ( Table 1). The average age at operation was 25.37 + 11.06 (10-52) years old. The mean pre-op scoliosis Cobb angle was 96.07 + 37.88 and the mean kyphosis angle was 94.43 + 57.08 .
Baseline data, such as sex, age, diagnosis, Cobb angle, fusion segments, osteotomy type was matched and there was no significant difference between the groups (P < .05). In the control group (Table 1); there were 35 patients (9 men and 26 women) with an average age of 26.14 + 8.73 years old, the mean pre-op Cobb angle of 104.83 + 28.86 , and the mean kyphosis angle of 95.38 + 37.47 .
There was no significant difference between 3DP group and control group in terms of the post-op scoliosis Cobb angle, kyphosis angle and correction rate (P > .05). However, the operation time was significantly shorter and blood loss was less in 3DP group (296.14 + 66.18 min vs. 329.43 + 67.16 min, 711.43 + 552.28 mL vs. 1322.29 + 828.23 mL, P < .05) ( Table 2).
Osteotomies were recorded according to Schwab spinal osteotomy classification system (Grade 1 to 6). 6 As shown in the Table 3, osteotomies were performed in most of the spine deformity correction surgeries (33/35 vs. 32/35). Statistical analysis revealed that there was no significant difference in the osteotomy distribution between groups. After dividing the osteotomies into posterior column osteotomy group (Grade 0-2) and 3-column osteotomy group (Grade 3-6), 3DP group exhibited a significantly higher rate of 3-column osteotomy than freehand group (30/35, 85.7% vs. 21/35, 60.0%.  The screw accuracy was higher in the 3DP group than in the control group, and the difference was statistically significant (w ¼ 11.26, P ¼ .02). The screw misplacement related complication rate, such as nerve root or spinal cord injury, was higher in the free-hand group (1/35 vs. 6/35), the difference was statistically significant (w ¼ 3.97, P ¼ .046) ( Table 4). The symptomatic misplaced screws were revised timely, all cases got satisfied recovery, except for 1 patient who was discharged with residual weakness in the left lower limb (Table 5).

Discussion
Three-dimensional printing, also known as additive manufacturing, is a kind of rapid prototyping technology. Medical 3DP is based on human anatomy to accurately make physical models of human organs and tissues. 7 The applications of 3DP for clinical purposes have grown rapidly over the past decade. 8 In the 1980s, it was first applied to engineering by Chuck Hull through computer modeling. 9 The medical application was developed in 1990, and a skull was printed based on the CT scan data. 10 Currently, 3DP plays an important role in orthopedics: in clinical teaching assistance, doctor-patient  communication, diagnosis of complex diseases, preoperative planning and surgical simulation training, surgical navigation, 3DP prosthesis implantation and bone tissue engineering. [11][12][13][14] Spinal deformity is 1 of the toughest challenges faced by spine surgeons and neurosurgeons. It is difficult for surgeons to evaluate and understand the exact pathological anatomical structure through traditional imaging examination, especially for congenital scoliosis caused by vertebra formation defect and failure of segmentation. However, with the emergence and development of 3DP technology, the situation has greatly improved. 1,15 Li et al 16 made 3DP models of 22 patients with cervical deformations, and simulated posterior cervical operations were conducted on the models to obtain accurate screw placement routes and angles. The screw accuracy was improved and no vertebral artery was injured. Sixteen severe scoliosis patients were treated with the help of 3DP technology, and the best screw placement angle and depth were obtained in the simulated surgery on the 3DP spine models before operation. 17 Li et al randomly divided 53 patients with scoliosis into the control and 3DP groups, and the statistical data showed that the  operation time was shorter and screw placement accuracy was better in the 3DP group.
It is challenging for surgeons to implant pedicle screws accurately in the deformed spine, which is usually accompanied by vertebra rotation, thin pedicle, hemivertebra, unsegmented vertebra, or multi-segment fused vertebra. Severe spine deformity cases treated by posterior pedicle screw fixation and fusion in our center were comprehensively compared between 3DP spine model-assisted group and free-hand group using PSM method in this study, and the results showed that pedicle screw placement was more accurate and operation time was much shorter with the guidance of the printed model. The operation time was shortened by 33 minutes on average, which would reduce surgery-related and anesthesia-related complications and improve the efficiency of operating room staff.
The severe spine deformity often requires spinal osteotomies to achieve good correction. The higher the osteotomy grade, the better the correction effect, but also comes with a higher risk. The osteotomy plan was made by the surgeon based on cognition of the spinal anatomy structure. It has been reported that preoperative familiarity of oncologic pathology based on 3D models in orthopedics helps to increase the accuracy of bone resection and to decrease operative time. 18 In this study, surgeons were more inclined to performed more advanced osteotomies with the help of 3DP spine models. In the meanwhile, neurological complications rate was reduced and operation time was shortened significantly.
Traditionally, CT scan and reconstraction images help us a lot to evaluate the spine deformity. Zheng et al reported that compared with reconstructed 3D-rendered images in preoperative planning, the utilizing of 3DP models could significantly improve surgical plan quality. 19 The authors postulated that the 3DP models may have improved the understanding of the anatomically complex sites of skeletal structure. 3DP models allowed surgeons to appreciate the structure and relations of  Left lower limb pain Yes Recovery the relevant anatomy much better than the visualization provided by 2-dimensional CT images conventionally used. 5 Inspection of these models also revealed structural abnormalities not appreciated on CT which altered the surgical approach in a significant number of cases. 20 Luo et al demonstrated that accuracy of the surgical technique using spinal 3D printing technology in patients with severe congenital scoliosis was higher than that of the free-hand technique, and it appeared to shorten operative time. 21 It is reported that the intraoperative navigation system or robotic assistant system have been used to enhance the accuracy of pedical screw insertion in the treatment of spine degenerative and some spine deformity disease, 22 and it turns out to be fesiable and effective. 23,24 However, in severe spine deformity surgery, the evidence was unpowered and insufficient. 25 Yang et al reported that 3D technology could reduce the misplacement rate, operating time and blood loss in patients with preoperative mean Cobb angle only >50 . 26 According to our experience, the application of robotic navigation technology in the operation of severe spinal deformity was not accurate and convenient enough.
The cost of 3DP full spine model varies from $400 to $500, which is much less than the surgical robot or surgical navigation. Thus, in the area of severe spine deformity surgery exhibiting the advantage of cost-effectiveness and reducing radiation exposure. 27,28 Currently, the progress of the surgeons' surgical skill learning curve mainly depends on clinical practice, cadaveric simulation, and digital imaging technology training, such as virtual reality and augmented reality. For surgeons in training, getting a chance to operate is competitive, especially in the developing countries where medical resources are inadequate, and this limits the training of surgeons. In addition, cadaveric training has significant limitations in cost, quality, and availability.
3DP technology has also been adopted for the surgeons' training and teaching purposes, whereby the design flexibility in terms of geometry and material properties (tissue density, hardness, flexibility) enables simulation of a range of clinical scenarios for surgical training 29,30 or anatomical learning 31 without the associated ethical and cost barriers, as well as anatomical variation, that can be present in cadaveric study. Especially for the complex spinal deformities, printed models can improve the junior doctors' understanding of the pathological structure of the spine. Zhao et al 32 trained junior surgeons by conducting simulated surgery with 3DP models. The accuracy of screw placement on the model reached 93.75% after 3-months of training. After 6 months of training, all junior surgeons could implant pedicle screw accurately and independently, which greatly reduced the training time. Besides, many studies have indicated that 3DP spine model could promote surgical education, team communication and patient understanding. [33][34][35][36] In conclusion, it is worthwile to apply the 3DP technology in severe spinal deformity surgery concerning the following advantages. Firstly, 3DP spine model improves the understanding of the anatomical structure of the surgical site and can be used for surgical rehearsal to accelerate the learning curve of surgeons and improve the quality of clinical education.
Secondly, it improves surgical safety by increasing screw accuracy and reducing fluoroscopy exposure. 16,17,37,38 Thirdly, it improves surgical efficiency by shortening the operation time. At last, 3D-printed spinal models make it easier for doctors to convince the patients and families about treatment plan, surgical complications, and rehabilitation program.
We suggest that 3DP models should be incorporated into the workflow in the surgical treatment of complex and severe spine deformity. It is believed that 3DP technology has great potential to become popular in the field of orthopedics. The medical 3DP industry may experience a period of rapid development.
We recognized that our study has limitations. First, it is a retrospective study with limited sample size, which will make our conclusions less convincing. Second, the process of grouping patients to different group is not random, leading to bias to some extent. Third, the correlation between the misplaced screws and the primary pathology of the deformity was not analyzed in the study, we will further our study to analyze the possible reasons of misplaced screws.

Conclusions
The study provided solid evidence that 3DP spine models can enhance surgeons' confidence in performing higher grade osteotomies and improve the safety and efficiency in severe spine deformity correction surgery. 3D printing technology has a good prospect in spinal deformity surgery.

Authors' Note
Aixing Pan and Hongtao Ding contributed equally to this paper. Aixing Pan: acquisition and analysis of data, drafting of the article, and statistical analysis; Hongtao Ding: acquisition and analysis of data, drafting of the article, and statistical analysis; Yong Hai: design, review and revision of the final manuscript, technical and administrative support, performed all the surgeries; Yuzeng Liu: statistical analysis and technical support, review and revision of the final manuscript. Junrui Jonathan Hai: acquisition and analysis of data; Peng Yin: statistical analysis and technical support; Bo Han: acquisition and analysis of data; all authors read and approved the final manuscript.
Informed consent was obtained from all individual participants included in the study. We have received the approval of the ethics committee of Beijing Chaoyang hospital, CCMU, and the study was performed in accordance with the Helsinki Declaration of 1964, and its later amendments. Informed consent was obtained from all individual participants included in the study.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by grants from the National Natural Science Foundation of China (No. 81 772 421).